NBA Clustering Report

Quinn Glovier

2024-02-21

Report

The goal of my data analysis was to find players that had high performance statistics compared to their salary on average. To classify by salary, I grouped the players into “Above Average”, “Average”, and “Below Average” based on the bottom 25% and top 75% thresholds for salaries. I chose to measure performance were Games Played, Games Started, Minutes Played, Field Goals, 3 Point Shots, 2 Point Shots, Free Throws, Total Rebounds, Assists, Turnovers, and Points Total.

When running the model, I grouped the data into three clusters, with Cluster 1 having the highest performing players and Cluster 3 having the lowest performing players. When evaluating the data with the “Elbow Chart” method and the “NBClusters” method, I found that 2 and 3 clusters were the most reccomended numbers, but 3 performed better.

When evaluating the results, the 3D graph was most helpful to me. I chose to review the players based on Minutes Played (shows how valuable of a player they are), 3 Point Shots (an offensive statistic), and Turnovers (a defensive statistic), so that the review process was more holistic and looked at different performance aspects. I selected below-average and average paid players in Cluster 1 to find high performing players with lower salaries. The players I picked are:

  1. Jakob Poeltl: This player had around 29 minutes of play time, which was shorter than some of the other high-performing players in the average salary group. He had 6 2-point shots and 1.6 turnovers, showing strong offensive capabilities.

  2. Cade Cunningham: This player was also within the average salary group and played on average around 32.6 minutes. He had 4.9 for 2 Point Shots and 3.7 Turnovers, which was one of the highest turnover scores, showing a strong defensive ability.

  3. Herbert Jones: This was one of, if not the best player in the below-average salary category, meaning he would probably be less expensive to recruit compared to the other two players. He had an average playing time of around 29.9 minutes, 2.3 2 Point Shots, and 1.8 Turnovers. Looking at his stats in the original dataset, Jones also had 3.8 Total Rebounds, and 9 Total Points, showing potential.

2D Graph Showing how Minutes Played and 3 Point Shots Correlate to Salary

3D Graph Showing How Minutes Played, 3 Point Shots, and Turnovers Affect Salary

“Elbow” Graph Reccomending the Number of Clusters to be Used

## Warning: Using `size` aesthetic for lines was deprecated in ggplot2 3.4.0.
## ℹ Please use `linewidth` instead.
## This warning is displayed once every 8 hours.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.

Graph Showing What Number Of Clusters is best for analyzing the data